Cloud Types Simulated by the NCEP GFS Model

Clouds and their interactions with radiation is recognized by the most recent IPCC (Intergovernmental Panel on Climate Change; 2001) report as the greatest uncertainty in future projections of climate, and, despite the considerable improvement of the physical realism of cloud representation in general circulation models, representation of clouds and their feedbacks is still the weakest component of current General circulation models (e.g., Senior and Mitchell 1993, Cess et al. 1996). In the last 5 to 10 years, cloud resolving models (CRMs) and single column models (SCMs) have become extensively used tools in evaluation and improvement of cloud representation in General circulation models (Randall et al. 1996). CRMs, because of their ability to explicitly simulate cloud-scale dynamics and meso-scale processes, are increasingly being used for understanding cloud processes (e.g. Chaboureau and Bechtold 2002, Kohler 1999, Krueger et al. 1995b, c). Recently, a high-resolution CRM has been embedded in a general circulation model to replace most of the physical parameterizations (Grabowski 2001). This multiscale modeling framework approach has been shown to produce the Madden-Julian Oscillation as well as higher-frequency tropical waves in a much more realistic manner than a general circulation model with a traditional cloud parameterization (Khairoutdinov and Randall 2001, Khairoutdinov et al. 2003). However, detailed evaluations of the ability of CRMs to represent the radiative effects of various cloud types have not been made. The usefulness of SCMs to evaluate/develop/improve cloud parameterizations in General circulation models has been improved partly by the increasing use of CRMs, and partly by the availability of more observations suitable for SCM studies produced by projects such as the Atmospheric Radiation Measurement program (ARM; Stokes and Schwartz 1994, Ackerman and Stokes 2003).